People already trust AI for everyday tasks, so why do most still refuse to let it near their bank account?

This explores how quickly people will let AI agents into their bank accounts and payments. The corpus suggests two different speeds. People are adopting AI in general faster than almost any technology before it. Letting AI handle money is much slower, because people's trust stops well short of their usage. A US survey from August 2025 put generative AI use at 54.6% three years after ChatGPT launched. That beats the personal computer (19.7%) and the internet (30.1%) at the same point in their histories Is generative AI adoption now outpacing the personal computer?. But familiarity doesn't carry over to money. A survey of 18,000 people in 15 countries found 45% think AI will change the world, yet only 29% would give an agent access to their bank account How willing are people to let AI control their finances?. A Harris Poll for Visa found the same pattern: 72% of US adults use AI assistants, but only 23% trust AI to make payments Does naming a payments brand increase trust in AI transactions?.

The more useful finding is what narrows that gap. In the Visa survey, 61% said they would trust Visa itself to handle payments started by an agent. People seem willing to delegate to a familiar financial brand that stands behind the agent, even when they won't delegate to the AI alone. An IMF analysis explains why that matters. It argues that agent payments should keep the AI's probabilistic judgment, meaning its best guesses rather than guaranteed outcomes, separate from the predictable, legally enforceable layers that authorize and settle the money Should AI agents make payment decisions independently?. Taken together, these suggest adoption will move at the speed of trusted infrastructure, not at the speed of AI capability. Infrastructure is still unsettled. Retailers are blocking agents because there is no standard way for an agent to identify itself or prove it is acting for someone What identity standards do shopping agents actually need?. Platforms may also split into open systems that accept any agent and closed systems that require their own, much as card networks did Will agentic commerce platforms adopt open or closed models?.

Two findings suggest that once someone starts, adoption can speed up sharply. Studies of autonomous shopping assistants find that perceived benefit drives acceptance, not a person's sense of control, and that scarcity strengthens the effect. Under scarcity, people hand over control when the payoff is obvious What drives shoppers to accept autonomous AI assistants?. Real behavior on the x402 agent-payment rails points the same way. The median time between funding an agent wallet and its first purchase fell from 73 days to 18 minutes, and the share of wallets funded again rose from 71% to 89% Are AI agent wallets becoming genuinely operational instead of experimental?. That is a small, early-adopter crypto population, but it shows what happens after the first hurdle: hesitation turns into habit.

The surprise is that the effects don't need mass adoption at all. McKinsey estimates that if only 10–20% of consumers let agents automatically move their cash to the best rate, deposit margins could shrink by 30–50 basis points. Banks have long profited from customers not bothering to switch, and that would break the habit Will AI agents break the deposit and credit card inertia?. The same minority could turn withdrawals from a weak bank, which used to take days, into a run that takes hours. That threatens individual banks rather than the whole system Could AI agents speed up runs on individual weak banks?. The 29% trust ceiling isn't evidence that change will be slow. It is already roughly the share of customers whose behavior could reshape banking economics.

The corpus has no forecasts or adoption curves specific to banking. What it offers is a pattern: trust stalls at around a quarter to a third of people, a brand-name intermediary can lift it, and use turns habitual fast once people start. The open question is whether the identity and settlement infrastructure arrives before or after that minority shows up.